基于HOS和Radon变换的图像盲复原算法改进

Improvement on the technology of image blind restoration based on HOS and Radon transform

针对一般图像复原算法中噪声在高频下放大的问题,提出了用高阶统计量对模糊系统进行辨识、估计出MA模型参数、再用反卷积算法估计出原始图像的复原算法.同时引入Radon变换将二维图像投影为某一角度下的一维投影,大大降低了算法的运算复杂度.仿真实验表明,该算法取得了良好的效果.

Aiming at the problem of noise＇s enlarging under the high frequency in general image restoration algorithm,an image blind restoration algorithm was put forward,which used higher statistics to identify the fuzzy system and estimate MA model parameters,and then estimate the former image with blind deconvolution.At the same time,two dimensional image can be projected as one projection from some angle through introducing Radon change to largely reduced the multiplication complex.Simulation results showed that the improved method was valulable.